Related Experiment Video
Updated: Sep 6, 2025
08:56
Terahertz Imaging and Characterization Protocol for Freshly Excised Breast Cancer Tumors
Published on: April 5, 2020
11.1K
Discriminating healthy from tumor tissue in breast lumpectomy specimens using deep learning-based hyperspectral
Lynn-Jade S Jong1,2,3, Naomi de Kruif4,3, Freija Geldof1,2
1Department of Surgery, Netherlands Cancer Institute, Plesmanlaan 121, 1066 CX Amsterdam, The Netherlands.
Biomedical Optics Express
|July 1, 2022
Summary
Hyperspectral imaging shows promise for improving breast-conserving surgery by accurately distinguishing tumor from healthy tissue. This technology can help surgeons achieve better resection margins, leading to improved patient outcomes.
Area of Science:
- Oncology
- Medical Imaging
- Computer Science
Background:
- Achieving adequate resection margins in breast-conserving surgery is difficult without real-time feedback.
- Accurate margin assessment is crucial for minimizing local recurrence and optimizing patient outcomes.
Purpose of the Study:
- To evaluate hyperspectral imaging (HSI) for differentiating tumor from healthy tissue in lumpectomy specimens.
- To assess the potential of HSI to provide intraoperative feedback for margin assessment.
Main Methods:
- Developed and evaluated three convolutional neural networks (CNNs) using a dataset from tissue slices.
- Fine-tuned the CNNs with lumpectomy data to predict tissue percentages on the resection surface.
- Utilized hyperspectral imaging to capture detailed spectral information from tissue samples.
Main Results:
- Achieved a Matthews Correlation Coefficient (MCC) of 0.92 on tissue slices, indicating high classification accuracy.
- Obtained a Root Mean Square Error (RMSE) of 9% when predicting tissue percentages on lumpectomy resection surfaces.
- Demonstrated successful discrimination between healthy and tumorous tissue using HSI.
Conclusions:
- Hyperspectral imaging holds significant potential for classifying resection margins in lumpectomy specimens.
- This technology could provide valuable intraoperative feedback, aiding surgeons in achieving complete tumor removal.
- Further development of HSI may enhance the efficacy of breast-conserving surgeries.
